CBIR

What is a new search concept? Watch this video and read this article (in Russian) to realize it fully.

Video retrieval (by means of CBIR techniques):

Viewdle – video search engine. Faces search in video-archives.
Video Google Demo. Search of similar scenes by image fragments.

TV news processing. Carnegie Mellon University.

Columbia University:

An object oriented video search engine. Search of similar scenes by image and sketch sample.
VisualSEEk – A joint spaital-feature image search engine. “The system finds the images that contain the most similar arrangements of similar regions.”

Image retrieval:

Tineye. Search by Image Query. It is a very convenient service with good quality performance. Recently Google has took a great interest in Tineye.
Picitup and it’s video on YouTube. They’ve bravely assumed the challenge “Can anyone do for images what Google does for text?”. And, we must admit, it is now one of the best visual search services. They offer CelebrityMatchUp (by faces), Visual Shopping (search similar products by colours, shapes, textures, brand and much more), Visual Search (first you search by text, and then precise by colour, simple shapes, simple categories).
GazoPa and the article about it (in Russian). It is a private search engine (only by invitation), you can search by example, sketch and outlines (by colour, texture, shape, spatial arrangement and worse by faces). Recently “GazoPa for iPhone” has been released.
QBIC (IBM’s Query By Image Content) and it’s State Hermitage Museum Project. You can search by sample and outline (by colour and set of colours with their percentage, by spatial organization of colours, by simple geometric figures and their spatial arrangement).

Photodate. Search people by photo and help in revealing the infringement of copyright.
Photo Finder on facebook. Scan photos and detect all faces appearing on them, recognize & auto tag (automatically match names to all faces), face galleries (search for friends and browse their face galleries – discover the photos you never saw before).
PolarRose. “Polar Rose detects and matches the faces in your photos – so you can easily name people and share photos with your Facebook friends”.

Like.com – Visual Shopping. You can search by brand, by category and by visual features: “Color Match”, “Shape Match”, “Pattern Match” (by means of texture), “Detail Search” (sub-image retrieval).
Etsy (”Shop by Color”). It is online shop of handmade goods. Unfortunately, you may choose only one colour, but the process of choosing colour is finely animated.

Visuvi Visual Search. Offers “Medical Image Search”, “Product Search” (Visual Shopping), search people (by faces) and products in social networks.
IRMA (Image Retrieval in Medical Applications). The student project.
NeuroInformatics (Mitre). NeuroImagery Retrieval & Visualization, including medical applications.

MFIRS (Multi-Features Image Retrieval System). Here user can choose the metrics and the method of images matching.
CIRES – Content Based Image REtrieval System and it’s new site. There are several individual classes with proper settings for every one (metrics, weights for different distances, colour space).
Behold. Offers very interesting opportunity – find images tagged with definite keywords that look like a picture of something (animal, beach, bird, boat, building, car, city, cloud, face and so on).
Collage. “The City of London Libraries, Archives and Guildhall Art Gallery is host to COLLAGE, an image database containing over 20,000 works of art from its collections.”
DEVA Group.
Viper – Multimedia Information Retrieval (Geneve University).
ImBrowse. A Browser for Large Image Databases.
Cbir.ru. Russian Massmedia Laboratory Project – search in image collections and videostreams.
Japan CBIR.

Retrievr. Interactive search by sketch, but results do not impress.
Search by sketch system. And video about it.

Viim and the article about it (in Russian). Do not impress. Service does not work correctly. According to the article, “big red ball” finds some red balls, “blue triangle” – some triangles.
Tiltomo and the article about it (in Russian). You can find similar photos by “theme” or by “color/texture”.
Riya. Assert that they can search by colour, texture and shape. Here we may search similar people (by faces), similar objects. Riya offers to use it’s “face recognition and text recognition, to search your personal photos”.
XCavator and the video about it. Search by example and sketch, by dominant colour and colour sets with their percentage, by spatial colour information, but first they find by keywords.
Piximilar. Visual similarity search for large image collections, it can be used in combination with keywords to refine searches on extremely large collections. “Piximilar’s visual similarity technology uses sophisticated algorithms to analyze hundreds of image attributes such as colour, shape, texture, luminosity, complexity, objects and regions.” Also here you can choose a set of colours with their percentage.
PixID. Identify editorial images in print and online publications. “PixID is a comprehensive, automated image monitoring service that uses advanced image identification algorithms to identify where your images are being used in print publications and the Internet.”
Recogmission, Picollator and their blog (in Russian). They use face recognition technology to search similar photos and organize photo archives. They assert that they can find objects on images (but nowadays only faces – it’s a great lack). They can cluster and classify image collection, search by visual example, reveal the infringement of copyright. There is also piFilter (the Web engine for automatic adult images detection). They offer Picollator.mobi for mobile search and funny variant of marking results (by tomato-icon).

Use Flickr results:

MUFIN. First search by keywords, then click on “Visually similar” link.
Flickr Color Selectr. We can choose the dominant colour.
Squared Circle Colr Pickr. We can also choose the dominant colour, the quantity of colours is much more.
Multicolr Search Lab. You can choose a set of colours with their percentage (by means of clicking the same colour several times).

Wang Research Group and the projects:

Wavelet Image Search Engine. Wavelet based images matching (by colour and texture). They also have wavelet based pornography elimination.
Simplicity and the same
Simplicity for Satellite Imagery and the same
CLUE (Cluster-based Image Retrieval)
Advancing Digital Imagery Technologies for Art and Cultural Heritages
Alip (Automatic Linguistic Indexing of Pictures)
Alipr (Automatic Photo Tagging and Visual Image Search). Fusion of visual search and text tagging – auto-tagging based on visual similarity.
Airplanes.net Image Similarity Search. Use Wang Simplicity technology.

Written by Sergey Egorov

No tags for this post.

Related posts

• No related posts.